are international undergraduates struggling academically? · undergraduates from the university’s...
TRANSCRIPT
60 Journal of International Students
Peer-Reviewed Article
ISSN: 2162-3104 Print/ ISSN: 2166-3750 Online
Volume 4, Issue 1 (2014), pp. 60-73
© Journal of International Students
http://jistudents.org/
Are International Undergraduates Struggling Academically?
Barry Fass-Holmes, PhD
University of California, San Diego (USA)
Allison A. Vaughn, PhD
San Diego State University (USA)
Abstract
Are international undergraduates struggling academically, and are their struggles due to weaknesses
in English as a second language? The present study showed that 1) at most 10% of these students in
three cohorts (ranging in size from N=322 to N=695) at an American west coast public university
struggled (quarterly grade point averages below C) in their university classes; 2) up to 63% of them
struggled with English (they were required to take a local community college’s English Composition
and/or English as a Second Language classes, and up to 42% earned course grades of D or F in
those classes); and 3) predictors shown to be statistically significant by hierarchical linear modeling
each accounted for less than 5% of the total variance (“small” effect sizes). These findings suggest
that only a minority of this university’s international undergraduates struggle in their university
classes even though a majority of them struggle with English.
Keywords: academic achievement, grade point average, international undergraduates, hierarchical
linear modeling
________________________________________________________________________________
“Our international undergraduates are struggling academically” is a generalization that some
administrators, advisors, and faculty have asserted recently at a west coast public university in the
United States. Administrators who have made this generalization additionally have stated that
international undergraduates in general are eligible for academic probation, disqualification, or
dismissal. Advisors who have made this generalization additionally have indicated that their
appointments with international undergraduates in general are about academic struggles. Faculty
who have made this generalization claim that international undergraduates’ English writing skills
generally are deficient.
This generalization drew attention and it has been repeated even though it was anecdotal
(the administrators, advisors, and faculty have not collected, statistically analyzed, or reported
relevant data to support their generalization, to our knowledge). Repeating this generalization
increases the likelihood of decisions that will be costly and/or ineffective due to the absence of
guiding data about the allegedly struggling undergraduates’ numbers and characteristics such as
applicant type (first-time freshman vs. transfer), class level, country of citizenship, gender, self-
reported immigration (I-94) status, or major department. For example, decisions could be made to
develop and implement specialized programs for all international undergraduates when, in reality,
only a small percentage needs or would benefit from such programs.
Journal of International Students 61
2014 http://jistudents.org Volume 4 • Issue 1
In the absence of guiding data, one possible explanation for the repetition of this generalization is
that an increase in the number of international undergraduates enrolled at the west coast public university
has resulted in a proportional increase in the number of ones struggling academically rather than that
academic struggles are a general characteristic of the university’s international undergraduates. For
example, the number of enrolled international undergraduates who struggle academically could change
from five out of 100 during one year to 10 out of 200 during the next and the proportion of struggling
international undergraduates would be stable rather than increasing. Such a scenario could give an
erroneous impression that the university’s international students in general are struggling academically
when, in fact, only a consistently small proportion of international undergraduates struggles from one
year to the next.
Many American universities have experienced increased enrollment of international
undergraduates, as shown by the Institute for International Education’s (IIE) recent annual Open
Doors snapshot survey (IIE, 2012a). This survey’s results showed that the total number of
international students enrolled in fall 2012 was higher than in fall 2011 at 61% (340) of
participating American universities. Moreover, academic year 2011–12 (AY1112) was the sixth
consecutive year in which IIE’s Open Doors report showed an increase in the total number of
international students in U.S. higher education—31% more international students studied at U.S.
colleges and universities in AY1112 than a decade ago. New internationals enrolling in 2011
increased 7% from 2010, and this increase was largely attributable to Chinese undergraduates
studying in the U.S. whose numbers were up 31% (IIE, 2012b).
An increase in the number of international undergraduates also has occurred at the
aforementioned west coast public university (cf. IIE, 2012a,b). This university is one of the Open
Doors snapshot survey’s 340 participants and its data in the snapshot survey show that it experienced
an increase in international undergraduates in fall 2012 compared to the previous fall (IIE, 2012a).
The number of new first-time international undergraduates (excluding transfers) at this university was
651 in fall 2012 which represents a 91.5% increase from the corresponding number for fall 2011—
340. It stands to reason that the number of international undergraduates who seek advising about
academic issues in AY1213 could be greater than in AY1112 simply because of the 91.5% increase
rather than because these students, as a whole, are struggling academically.
To address administrators’, advisors’, and faculty’s concern that the west coast public
university’s international undergraduates are struggling academically, the present study was conducted
with the primary goal of determining what percentage of incoming international undergraduates do
struggle academically as reflected by their grade point averages (GPA). Another goal was to evaluate
how this percentage has changed relative to the number of these students admitted to this university; did
it increase, stay the same, or decrease between years? To address administrators’, advisors’, and faculty’s
concern that international undergraduates’ academic struggles are due to English weaknesses, the present
study also evaluated the extent to which these students at the west coast public university showed
evidence of struggling with English. Lastly, the present study investigated the potential role of other
variables (such as applicant type, class level, country of citizenship, gender, self-reported immigration (I-
94) status, or major department) in these students’ academic struggles.
Literature Review
Our internet-based literature search found very few published articles on international
undergraduates’ academic struggles and/or GPAs while attending American universities (described
below). Consequently, this literature review focuses primarily on recent developments that provide
context for the present study’s goal to determine what percentage of incoming international
undergraduates do struggle academically, and whether the struggles could be attributed to English
language weaknesses or other characteristics of these students.
62 Journal of International Students
State government funding for American public universities has decreased since 2008. It
declined nationwide by 7.5% in fiscal year 2010–11 and by an additional 0.4% in the current fiscal
year (Kelderman, 2013). Total funding for public universities was 10.8% lower this year than
before the 2008 economic downturn according to data released by Illinois State University’s
Grapevine Project (2013).
To compensate for shrinking state government support, American public universities have
increased their enrollment of international undergraduates as evidenced by IIE’s (2012b) Open
Doors data. A likely explanation for this increase is that international undergraduates pay non-
resident tuition fees that are higher than the resident fees paid by in-state undergraduates. However,
the financial benefit of admitting an increasing number of international undergraduates is offset by
the financial cost of providing additional programs and services to meet these students’ special
needs regarding immigration regulations, career, academics, communication, culture, personal
issues, and discriminatory treatment (Hanassab & Tidwell, 2002).
One approach that admissions offices could use to minimize costs associated with programs
and services for international undergraduates’ special needs would be to accurately predict each
applicant’s likelihood of academic success. Accordingly, research on international students
historically has focused on identifying English proficiency-related indicators for evaluating
applicants. Early studies measured the predictive power of international undergraduates’ scores on
the Test of English as a Foreign Language (TOEFL), and these studies produced inconsistent results
(reviewed by Graham, 1987; Johnson, 1988; Stoynoff, 1997)—some showed that TOEFL scores
were positively correlated with international undergraduates’ GPA at American universities while
others showed no correlation. A more recent study on undergraduates from China, Taiwan, Kuwait,
Saudi Arabia, and the United Arab Emirates at a large state university found that a passing TOEFL
score was no better at predicting academic success than a passing grade in English as a Second
Language (ESL) class (Chen & Sun, 2006).
Indicators unrelated to English proficiency also have been studied for their potential
usefulness in predicting international applicants’ future academic success. In particular, self-
confidence and high school class rank (House, 2000), availability of a strong support person (Boyer
& Sedlacek, 1988), self-efficacy, optimism, and academic expectations (Chemers, Hu & Garcia,
2001) were found to be significantly and positively related to international undergraduates’
academic success in American universities. Previous studies have not evaluated the predictive
power of international undergraduates’ applicant type (first-time freshman vs. transfer), class level,
country of citizenship, gender, self-reported immigration (I-94) status, or major department (cf.
Hanassab & Tidwell, 2002; Nelson, Nelson & Malone, 2004; Ren & Hagedorn, 2012), to our
knowledge.
Methods
Demographic and academic achievement data for three cohorts of the west coast public university’s
incoming (new freshman and transfer) international undergraduates were extracted from the
university’s student information system data tables. For the sake of consistency with U.S.
Government regulations’ definition of non-immigrant international students (U.S. Department of
State, n.d.), the three cohorts of non-resident aliens excluded domestic undergraduates and
additionally excluded amnesty-seekers, applicants for permanent residency, asylees, permanent
residents, refugees, and undocumented individuals. These cohorts of international undergraduates
began attending the university in fall 2009 (FA09), fall 2010 (FA10), and fall 2011 (FA11)
respectively. The reasons for using these particular cohorts in the present study were 1) these fall
quarters were the most recent ones for which grade point averages (GPAs) were available at the time
Journal of International Students 63
2014 http://jistudents.org Volume 4 • Issue 1
this study was conducted, and 2) the international undergraduates who entered in FA09 represented a
“baseline” year preceding the ones who entered in FA10 for whom the Admissions Office piloted
enhanced recruitment procedures (virtual outreach tools) targeting international applicants, and the
ones who entered in FA11 were the first for whom these procedures were fully implemented. The
Admissions Office’s enhanced recruitment procedures (attending virtual fairs, distributing recruitment
materials electronically, collecting information from overseas high schools, and attending college fairs
in the U.S. that provided opportunities for direct contact with foreign high school officials) were
intended to increase the “yield” of international applicants accepting the offer of admission, and were
in response to the university’s shortfall of state funding. Consequently, the university’s administrators
had a strong need for and interest in academic achievement (GPA) data for these three cohorts of
international undergraduates which the present study was designed to fulfill.
To extract longitudinal demographic and academic data plus SAT and TOEFL scores
(described below) for the three cohorts of the west coast public university’s international
undergraduates from the university’s student information system databases, structured query
language (SQL) programs were written and executed. The SQL programs also extracted each
international undergraduate’s unique campus ID and first and last names to ensure that every
academic quarter’s and year’s data were correctly organized within appropriate records in data files
for the statistical analyses described below. To protect confidentiality, prior to the statistical
analyses, unique dummy IDs were assigned to each international undergraduate in the extraction
files, then the files were duplicated and personally identifiable data (IDs; first and last names) were
permanently deleted. This procedure was IRB approved.
The statistical procedures performed on these data files to determine what percentage of
international undergraduates struggled academically and the degree to which their struggles were
related to weaknesses in English language included descriptive and correlational analyses (using
StatView software). Authentic 0.00 GPAs (e.g., all course grades of F in an academic quarter’s
classes) were included, artifactual ones (e.g., all “Pass” in an academic quarter’s classes) were
excluded. GPAs below 2.0 (C) are considered “struggling” and ones at or above 3.5 (between B+ and
A-) are considered “excelling” at the west coast public university.
To evaluate the potential role of other variables (such as applicant type, class level, and
major department) in the three cohorts of international undergraduates’ academic struggles, we also
performed hierarchical linear modeling (HLM; using HLM 6 software; Raudenbush & Bryk, 2002)
analyses and computed the effect sizes of all predictor variables. Each of our models was run with
only one predictor at a time, and most of our predictor variables were dummy-coded. These
included applicant type (first-time freshman vs. transfer), country of citizenship (China vs. all
others; see explanation below), gender (male vs. female), self-reported immigration (I-94) status
(F1 vs. all others), and major department (math vs. all others, economics vs. all others, engineering
vs. all others, computer science vs. all others). We selected departments where weakness in English
language might play more (or less) of a role in international students’ academic struggles. Class
level (freshman, sophomore, junior, senior) was left as an ordinal predictor variable. The dependent
variable in these analyses was quarterly GPAs.
The primary reason for using HLM to evaluate the potential role of the above variables in
international undergraduates’ academic struggles rather than using other statistical tests (such as
analysis of variance) which involve ordinary least squares estimation is that HLM provides better
parameter estimates when data are hierarchically structured (Osborne, 2000; Raudenbush & Bryk,
2002). An additional reason is that HLM, in conjunction with maximum likelihood estimation,
uniquely handles missing data, thereby precluding the need to exclude students who lack at least
one academic quarter’s data (due to leave of absence, withdrawal, etc.), and precluding the need to
use imputations (which can be controversial; Little & Rubin, 2002).
64 Journal of International Students
The HLM analyses also addressed one additional question—are permanent residents (PR)
and/or undocumented (OT) undergraduates the students with academic struggles rather than
internationals? To address this question, we included PR and OT undergraduates (but not
domestics) in another set of record extractions from the university’s student information system
similar to the three international cohorts’ records described above and included them in the HLM
analyses.
Results
Descriptive Analyses
The total numbers of new international undergraduates—first-time freshmen (NFRS) and transfers
(TRAN)—registering for classes at the west coast public university have increased over recent
years as shown in Figure 1. The FA10 cohort is 36.4% larger than the FA09, the FA11 cohort is
56.4% larger than the FA10, and the FA12 cohort is 44.0% larger than the FA11 (the FA 12 cohort
is included only in Figure 1 to demonstrate the trend; GPA data for this cohort were not yet
available at the time of the present study).
Figure 1. The number (above each
bar) of and year-to-year percentage
change (within the last three bars)
in international undergraduates
registered for classes at the west
coast public university in the
present study has increased
consistently between fall 2009
(FA09) and fall 2012 (FA12).
Abbreviations: FA10=fall 2010;
FA11=fall 2011
Demographic data for the
three cohorts are presented in Table
1. This study focused on comparing
international undergraduates from
China with counterparts from other
countries (including but not limited
to Hong Kong, India, South Korea,
and Taiwan) because 1) China has
become the predominant country of
citizenship for international
undergraduates attending the west
coast public university in the
present study (as reflected by the dramatic increase in Chinese undergraduates between FA09 and
FA11 compared with the corresponding increase in all others; Table 1), and 2) Chinese
undergraduates anecdotally are perceived at this university as the ones most likely to have English
weaknesses. International undergraduates majoring in the departments listed in Table 1 anecdotally
are perceived on campus as the ones most likely to have English weaknesses (these majors
allegedly are less sensitive to English weaknesses, hence international undergraduates with English
weakness prefer them according to the anecdotes).
Journal of International Students 65
2014 http://jistudents.org Volume 4 • Issue 1
Table 1
Demographic Characteristics of the Three International Undergraduate Cohorts Admitted in Fall
2009 (FA09), Fall 2010 (FA10), or Fall 2011 (FA11)
Demographic
FA09 FA10 FA11 % increase from
FA09 to FA11
Applicant type new freshmen (NFRS) 69 155 337 388.4
transfer students (TRAN) 253 286 358 41.5
Total 322 441 695 115.8
Class level freshman 62 147 323 421.0
sophomore 15 10 18 20.0
junior 243 283 350 44.0
senior 2 1 4 100.0
Total 322 441 695 115.8
Country China 35 110 310 785.7
all othersa
287 331 385 34.1
Total 322 441 695 115.8
Department Computer Science and Engineering 11 17 44 300.0
Economics 137 172 251 83.2
Electrical and Computer Engineering 8 14 36 350.0
Mathematics 4 15 47 1,075.0
all others 162 223 317 95.7
Total 322 441 695 115.8
Gender female 157 214 303 93.0
male 165 227 392 137.6
Total 322 441 695 115.8
Note. aAll other countries include (but are not limited to) Hong Kong, India, South Korea, Taiwan,
plus others for which fewer than three students attended the university.
Academic achievement data for the three cohorts are shown in Figures 2, 3, and 4. Figure 2
contains each of the three cohorts’ mean GPAs for each of the three academic quarters (fall, winter,
and spring) of the year in which they entered the west coast public university (AY0910 for the
FA09 cohort, and so on). The international undergraduates in these cohorts, in general, earned a
GPA between 2.8 and 3.1 (roughly between B- and B). Figure 3 contains the percentage of each of
the three cohorts that earned GPAs below 2.0 (i.e., struggled academically) and Figure 4 contains
the corresponding percentages that earned GPAs at or above 3.5 (i.e., excelled academically). The
percentages show that only a tenth or fewer of international undergraduates in these cohorts
struggled academically and a larger minority (between a quarter and a third) excelled academically.
66 Journal of International Students
Figure 2. The mean quarterly GPA
(above each bar) ranged from 2.86
(between B- and B) to 3.07 (just
above B) for each of the three
international undergraduate cohorts
registered for classes at the west coast
public university in the present study.
These GPAs were higher than the
value (below 2.0) considered to be
“academically struggling” at this
university.
Abbreviations:
FA09=fall2009; WI10=winter 2010;
SP10=spring 2010; FA10=fall 2010;
WI11=winter 2011; SP11=spring 2011;
FA11=fall 2011; WI12=winter 2012;
SP12=spring 2012.
Figure 3. The percentage of each of
the three international undergraduate
cohorts in the present study that had a
quarterly GPA below 2.0 (“struggled
academically”) ranged from 6.8 to
10.2. These percentages are less than
what would be expected if GPAs were
distributed normally.
Abbreviations:
FA09=fall 2009; WI10=winter 2010;
SP10=spring 2010; FA10=fall 2010;
WI11=winter 2011; SP11=spring 2011;
FA11=fall 2011; WI12=winter 2012;
SP12=spring 2012.
Figure 4. The percentage of each of
the three international undergraduate
cohorts in the present study that had a
quarterly GPA at or above 3.5
(“excelled academically”) ranged
from 23.1 to 36.8. These percentages
exceeded the corresponding
percentages that struggled
academically.
Abbreviations:
FA09=fall 2009; WI10=winter 2010;
SP10=spring 2010; FA10=fall 2010;
WI11=winter 2011; SP11=spring 2011;
FA11=fall 2011; WI12=winter 2012;
SP12=spring 2012.
Journal of International Students 67
2014 http://jistudents.org Volume 4 • Issue 1
While the number of the west coast public university’s international undergraduates who earned
GPAs below 2.0 (i.e., struggled academically) amounted to a tenth or less, the percentage with
demonstrable English weaknesses was higher as evidenced by the data in Figure 5. This figure shows
the percentage of each of the three cohorts’ F1 (self-reported I-94 status) NFRS who were required to
attend a local community college’s English Composition and/or ESL class (this requirement is part of
the university’s entry level writing requirement for undergraduates). This percentage increased across
the cohorts (reaching almost two-thirds of the FA11 cohort). Further evidence of these
undergraduates’ English weaknesses appears in Table 2. This table includes the percentages of the
three cohorts’ F1 (self-reported I-94 status) NFRS who earned a D or F in these community college
writing and ESL classes. Although this percentage increased across the cohorts (reaching more than a
third of the FA11 cohort), these students’ mean FA GPA in their university classes was between 3.24
and 3.33 (roughly between B and B+). Less than one tenth of them earned GPAs below 2.0 (Table 2).
Figure 5. The percentage of international new freshman undergraduates (NFRS) in F1 I-94 status who were
required to take a local community college’s English Composition or English as a Second Language class
has increased during the past three fall quarters at the west coast public university in the present study.
Table 2
Numbers and Percentages of International (F1 I-94 Status) New Freshmen (NFRS)
Cohort # F1
NFRS
# (%) required
to take
CC classes
% who earned
D or F in CC
classes
Mean FA GPA in
university classes
of F1 NFRS required
to take CC classes
# (%) who earned GPA
<2.0 in university’s FA
classes
FA09 48 18 (37.5) 5.6 3.33 0 (0.0)
FA10 121 57 (47.1) 10.5 3.24 3 (5.3)
FA11 319 200 (62.7) 42.0 3.31 13 (6.5)
Note. Abbreviations: CC=community college; FA=fall quarter
68 Journal of International Students
Correlation Analyses
To evaluate the potential role of English weaknesses (especially writing) in international
undergraduates’ academic struggles, we performed correlation analyses between these students’
scores on English proficiency tests (TOEFL, Scholastic Aptitude Test (SAT), or the west coast
public university’s writing exam) and corresponding AY GPAs (AY0910 for the FA09 cohort, and
so on). Significant positive correlation coefficients are indicative that the test score included in the
analysis is a predictor of the quarterly GPA included in the analysis—the higher the test score, the
higher the quarterly GPA; the lower the test score, the lower the quarterly GPA.
Table 3 shows the statistically significant results of these correlational analyses for F1
NFRS only (these are the international undergraduates of particular concern/interest for campus
policy and programming purposes). SAT math was the only one of the aforementioned scores that
consistently predicted the cohorts’ mean GPAs; however, these significant SAT math correlations (r
between 0.15 and 0.3) fall within the range of “small” (magnitude) effect sizes (Cohen, 1988). None
of the aforementioned English proficiency tests consistently correlated with mean GPAs; the only
significant correlations were between the FA10 cohort’s SAT writing scores and FA10 GPAs, and
between the FA10 cohort’s writing exam scores and FA10 GPAs (Table 3).
Table 3
Statistically Significant Predictors of Quarterly GPAs from the Correlation Analyses
Cohort Predictor Quarter r df p
FA09 SAT math FA09 0.27 48 0.05
SAT math WI10 0.30 48 0.05
FA10 SAT math FA10 0.23 111 <.05
SAT math WI11 0.23 109 <.05
SAT math SP11 0.26 107 <.01
SAT writing FA10 0.32 111 <.01
writing exam FA10 0.29 109 <.01
FA11 SAT math FA11 0.23 300 <.0001
SAT math WI12 0.21 294 0.0003
SAT math SP12 0.15 295 0.01
Note. Abbreviations: r = correlation coefficient, df = degrees of freedom, p = probability
(significance level)
HLM Analyses
The data extracted for the HLM analyses yielded 982 undergraduates’ records (including PRs and OTs,
excluding domestics) for the FA09 cohort, and these undergraduates had a total of 2,894 term GPAs used in
the HLM analyses. Consequently, on average, each FA09 undergraduate in the HLM analyses had about
three academic quarters (M = 2.95) of GPA data. The corresponding values for the FA10 cohort were 1,151
Journal of International Students 69
2014 http://jistudents.org Volume 4 • Issue 1
undergraduates, 3,396 term GPAs, and an average of almost three academic quarters (M = 2.95) of GPA
data per student. The FA11 cohort’s values were 1,274 undergraduates, 3,752 term GPAs, and an average
of almost three academic quarters (M = 2.95) of GPA data per student.
In the current study, all of our models were run with only one predictor at a time and
primary analyses were conducted on two-level models. GPA (the dependent variable) is considered
a lower-level (level-1 or time-varying) variable, whereas the various predictors are considered
higher-level (level-2 or stable) variables. These included applicant type (NFRS; TRAN), class level
(freshman, sophomore, junior, and senior), country of citizenship (i.e., China), gender, I-94 status
(i.e., F1), and major department (e.g., Economics, Mathematics). An example model is shown
below.
Level 1: GPA = β0j + rij
Level 2: β0j = γ00 + γ01(Class Level) + u0j
For each HLM model, only statistically significant results are presented in Table 4. This table
shows that the FA09 cohort’s class level was a significant predictor of GPA—as class level increased,
GPA decreased (indicated by the negative regression coefficient). Applicant type, country of citizenship,
and major department were all significant predictors of GPA—NFRS, Chinese students, and
Mathematics majors had higher GPAs compared to TRAN, non-Chinese students, and all other majors,
respectively. Gender and self-reported immigration (I-94) status were not significant predictors of
quarterly GPAs.
Table 4
Statistically Significant Predictors of GPA from the Hierarchical Linear Modeling Analyses
Cohort Predictor β SE β p sr2
FA09 Applicant type -0.129 0.039 0.001 0.012
Class level -0.039 0.020 0.051 0.003
Country of origin 0.142 0.051 0.006 0.008
Department 0.344 0.135 0.012 0.004
FA10 Applicant type -0.096 0.036 0.008 0.007
Country of origin 0.182 0.042 <.001 0.019
Department 0.206 0.085 0.015 0.004
FA11 Applicant type -0.216 0.039 <.001 0.031
Class level -0.083 0.020 <.001 0.018
Country of origin 0.270 0.041 <.001 0.043
Department 0.214 0.102 0.036 0.005
I-94 status -0.109 0.039 0.006 0.008
Note. Abbreviations: β = HLM model’s regression coefficient; SE β = standard error of the
regression coefficient; p = probability (significance) level; sr2
= semi-partial correlation squared
(analogous to R2), the proportion of variance accounted for in linear regression.
70 Journal of International Students
The FA10 cohort’s results were the same with one exception—class level was not a significant predictor
of GPA. The FA11 cohort’s results also were the same as the FA09’s with one additional predictor—
self-reported immigration (I-94) status was significant (i.e., F1 visa holders had higher GPAs than the
PRs, OTs, and other visa status immigrant students).With regard to the aforementioned additional
question, are PR and/or OT undergraduates the students with academic struggles rather than
internationals, the above results are consistent with the view that the PRs and OTs probably are not
the students that administrators, advisors, and faculty at the west coast public university have
referred to when repeating the generalization that “international students are struggling
academically.”
Although the HLM analyses revealed that the above predictors were statistically significant,
each one accounted for less than five percent of the total variance. They consequently fall within the
range of “small” (magnitude) effect sizes (Cohen, 1988).
Discussion
The current study’s primary goals were to 1) systematically evaluate whether international
undergraduates at the west coast public university in this study actually are struggling academically,
and 2) statistically determine whether the struggles could be attributed to English language
weaknesses and/or related to other variables. Regarding the first goal, the present findings do not
support anecdotal reports that international students at this university in general are struggling
academically; at most a tenth of the FA09, FA10, and FA11 cohorts earned GPAs that met the
campus definition of “struggling”—below 2.0 (C or “average”). However, regarding the second goal,
the present findings do indicate that an increasing percentage of F1 NFRS has English weaknesses as
indicated by required participation and course grades in community college English classes. We
conclude from these findings that the majority of the west coast public university’s international
undergraduates are not struggling academically, that the struggling ones comprise only a small
percentage of these students, and that the three cohorts generally succeeded in their university classes
despite evidence of struggling with English (shown in Table 2).
If the percentage of academically struggling international undergraduates is small, then what
accounts for anecdotal reports that these students generally are struggling? This west coast public
university historically has provided strong support to its international undergraduates through a wide
range of programs and services (e.g., orientations; academic and immigration advising; one-on-one
English tutoring; social and cultural events; etc.), therefore anecdotal reports of academic struggles
cannot readily be attributed to a lack of assistance and/or support. Instead, one possible explanation is
that the reports are a side effect of annual increases in international undergraduates entering this
university. It stands to reason that as more international undergraduates attend any American
university, the number who struggle academically likely will increase also. Importantly, however, the
percentages of international undergraduates who struggle academically were shown to remain
relatively stable (one tenth or less) in the present study, and these percentages are less than what
would be expected if GPAs were distributed normally (approximated a Bell curve). An alternative
explanation is that the struggling undergraduates include immigrant (applicants for permanent
residency, amnesty-seekers, asylees, permanent residents, refugees, and/or undocumented students)
students rather than or in addition to non-immigrant (international) students. Because we do not have
access to the identity and immigration (I-94) status of the undergraduates referenced in the anecdotal
reports, we cannot presently evaluate this alternative explanation. A third possible explanation is that
continuing, rather than new, international undergraduates are the ones who struggle academically. To
address this possibility, the present study will follow these three NFRS and TRAN cohorts’ academic
performance in their continuing years at the west coast public university.
Journal of International Students 71
2014 http://jistudents.org Volume 4 • Issue 1
What variables account for the academic struggles of the small percentages of international
undergraduates described in the present report? To evaluate the possibility that English language
weaknesses (especially in writing) account for these students’ struggles, we statistically analyzed
three cohorts’ required participation in community college English classes (English Composition;
ESL) and their scores on standardized English proficiency tests. The percentage of F1 NFRS who
were required to attend the community college English classes increased between the three cohorts,
reaching almost two thirds of the FA11 cohort. Consistent with anecdotal reports, the percentage
who struggled in these community college classes (i.e., earned course grades of D or F) increased
between the cohorts, reaching 42% of the FA11 cohort (Table 2). Although these international
undergraduates struggled with English writing as evidenced by their participation and course grades
in the community college English classes, their mean GPA in university classes during their first
(FA) quarter was between B- and B, and less than one tenth of their mean GPAs was lower than 2.0.
How could these international undergraduates have English weaknesses (as evidenced by
their performance in the community college classes) while simultaneously succeeding academically
in their university classes (as evidenced by their GPAs)? One possible explanation is that the
community college classes (English Composition; ESL) are sensitive to English weaknesses while
the university classes (e.g., Computer Science; Economics; Engineering; Mathematics) are not.
Another is that the community college instructors grade these students more strictly on English
(grammar, spelling, etc.) while university instructors grade less strictly on English (instead focusing
on whether the students show evidence of mastering class concepts). An additional possibility is
that these students invest more time and energy on their university classes than on the community
college classes. Further research will be needed to evaluate these possible explanations.
If almost two thirds of the west coast public university’s international undergraduates do
have English weaknesses, how could they qualify for admission to that university? International
applicants are required to submit TOEFL and SAT scores; applicants who take either of these
exams multiple times are considered on the basis of their highest total TOEFL and highest
combined SAT scores. One possibility is that international applicants attend courses in their home
countries that teach to these tests without improving the applicants’ English proficiency. Another is
that applicants who take these tests multiple times show a practice effect. A third possibility is that
at least some applicants cheat on TOEFL and/or SAT. Additional research is needed to evaluate
these three possibilities. Regardless, the present study’s non-significant correlation coefficients
between TOEFL or SAT reading and writing scores and quarterly GPAs suggest that while these
standardized tests might be useful for admissions purposes, they are not strong predictors of
international undergraduates’ academic struggling in their first year of university classes.
The present study’s finding that TOEFL scores were not correlated with quarterly GPAs is
inconsistent with a previous study’s (Stoynoff, 1997) finding that TOEFL scores were significantly
positively correlated with GPA. One possible explanation for the difference between these two
studies’ findings is that the 1990s’ TOEFL differed from the modern TOEFL in at least two
potentially important and related dimensions—range of the scoring scale, and mode of
administration (paper-based vs. online; Educational Testing Service, 2005). Another possibility
relates to a difference in statistical methodology—Stoynoff (1997) used a statistical correction to
account for the narrow range of international applicants' TOEFL (and, by extension, SAT) scores
due to universities' minimum requirements, we did not. In either case, Stoynoff (1997) described the
correlation between TOEFL scores and international undergraduates’ GPA as “modest” and thus,
combined with our finding, it suggests that TOEFL’s utility in predicting international
undergraduates’ GPAs at American universities is limited.
The present study is the first, to our knowledge, using HLM to investigate what variables
affect international undergraduates’ academic achievement while attending American universities (cf.
72 Journal of International Students
Li, Chen & Duanmu, 2010). Our HLM analyses included additional variables—applicant type, class
level, country of citizenship, gender, immigration (I-94) status, and major department—and some of
them were shown to be significant predictors for all three cohorts’ GPAs. However, these significant
predictors accounted for a low percentage of the total variability; they would be considered “small”
effect sizes (Cohen, 1988). Consequently, our findings might not be sufficiently compelling for use in
decision making about implementing or changing policies and programs for addressing international
undergraduates’ academic struggles. We instead recommend the development and implementation of
a diagnostic instrument that correlates highly with quarterly mean GPAs for use in preemptively
identifying which international undergraduates to target for additional support.
In conclusion, the present results taken together suggest that policies and programs intended
to support newly admitted international undergraduates with weak English skills would be most
cost effective if they were implemented for such students with demonstrable evidence of academic
struggles and/or English weaknesses rather than for all incoming international undergraduates.
References
Boyer, S., & Sedlacek, W. (1988). Non-cognitive predictors of academic success for international
students: A longitudinal study. Journal of College Student Development, 29(3), 218–223.
Chemers, M. M., Hu, L. & Garcia, B. F. (2001). Academic self-efficacy and first-year college
student performance and adjustment. Journal of Educational Psychology, 93(1), 55–64.
Chen, Y., & Sun, C. (2006). Language proficiency and academic performance. Proceedings of the
11th Conference of Pan-Pacific Association of Applied Linguistics, Korea, 11, 58–72.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.
Educational Testing Service (2005). TOEFL® internet-based test score comparison tables. Retrieved from
http://www.ets.org/Media/Tests/TOEFL/pdf/TOEFL_iBT_Score_Comparison_Tables.pdf
Graham, J.G. (1987). English language proficiency and the prediction of academic success. TESOL
Quarterly, 21(3), 505–521.
Grapevine Project (2013). One-year (FY12-FY13), two-year (FY11-FY13), and five-year (FY08-
FY13) percent changes in state fiscal support for higher education, by state and by source of
fiscal support. Retrieved from http://grapevine.illinoisstate.edu/tables/FY13/Table2_FY13.pdf
Hanassab, S., & Tidwell, R. (2002). International students in higher education: identification of
needs and implications for policy and practice. Journal of Studies in International Education,
6(4), 305–322.
House, J. D. (2000). Relationships between self-beliefs, academic background, and achievement of
undergraduate students in health sciences majors. International Journal of Instructional Media,
27(4), 427–438.
IIE (2012a). U.S. campuses report on impact of international students from China; engagement with
Brazil; students affected by Arab Spring events. Retrieved from http://www.iie.org/Who-We-
Are/News-and-Events/Press-Center/Press-Releases/2012/11-13-2012-Open-Doors-Fall-Survey-
International-Students
IIE (2012b). Open Doors 2012: international student enrollment increased by 6 percent. Retrieved
from http://www.iie.org/en/Who-We-Are/News-and-Events/Press-Center/Press-
Releases/2012/11-13-2012-Open-Doors-International-Students
Johnson, P. (1988). English language proficiency and academic performance of undergraduate
international students. TESOL Quarterly, 22(1), 164–168.
Kelderman, E. (2013). State spending on higher education rebounds in most states after years of
decline. The Chronicle of Higher Education. Retrieved from http://chronicle.com/article/State-
Spending-on-Higher/136745/
Journal of International Students 73
2014 http://jistudents.org Volume 4 • Issue 1
Li, G., Chen, W., & Duanmu, J-L. (2010). Determinants of international students' academic
performance: a comparison between Chinese and other international students. Journal of Studies
in International Education, 14(4), 389–405.
Little, R. J. A., & Rubin, D. B. (2002). Statistical analysis with missing data (2nd ed.). New York,
NY: John Wiley.
Nelson, C. V., Nelson, J. S., & Malone, B. G. (2004). Predicting success of international graduate
students in an American university. College and University Journal, 80(1), 19–27.
Osborne, J.W. (2000). Advantages of hierarchical linear modeling. Practical Assessment, Research,
and Evaluation, 7(1), 1–3.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models. Newbury Park, CA: Sage.
Ren, J., & Hagedorn. L. S. (2012). International graduate students’ academic performance: What
are the influencing factors? Journal of International Students, 2(2), 135–143.
Stoynoff, S. (1997). Factors associated with international students’ academic achievement. Journal
of Instructional Psychology, 24(1), 56–68.
U.S. Department of State. (n.d.). Student visas. Retrieved from http://travel.state.gov/
visa/temp/types/types_1268.html
_______________
About the Authors:
Barry Fass-Holmes received his Ph.D. in Psychology from Clark University. He currently is
SEVIS and Analytical Studies Coordinator for the International Students & Programs Office in the
International Center at the University of California, San Diego. His research focuses on
international students’ academic achievement. He thanks Lynn C. Anderson, Dulce Amor L.
Dorado, Dr. Kirk Simmons, and Prof. dr. J.W.M. de Wit for their helpful suggestions. His email
address is [email protected].
Allison A. Vaughn received her Ph.D. in Psychology from the University of Utah. She currently is
an assistant professor in the Psychology Department at San Diego State University. Her research
interests are social relationships (friends, significant others, work colleagues), mental health
(anxiety, depression), and physical health (cardiovascular functioning like blood pressure and heart
rate). Her email address is [email protected].